In the last decade, artificial intelligence (AI) has increasingly been applied to help solve applied ecology problems. Partially observable Markov decision processes (POMDPs) are one such example. POMDPs have been applied in conservation, applied ecology and natural resource management to solve problems such as deciding when to stop managing or surveying threatened species that are difficult to detect. POMDP solvers are useful to find optimal sequential decisions under imperfect detection. However, POMDPs remain inaccessible to most applied ecologists.
We present the shiny r package smsPOMDP that solves the problem of ‘When to stop managing or surveying cryptic threatened species?’ (Chadès et al., 2008). We developed this package to address a common and challenging problem faced by conservation managers.
In artificial intelligence, POMDPs are acknowledged as the Swiss army knife of decision models. However, POMDP’s application in applied ecology remains seldom despite repeated evidence of their flexibility. Our package smsPOMDP is fast and provides an entry point to further develop POMDP apps, contributing to further uptake of AI research to solve ecological problems.
Pascal, L, Memarzadeh, M, Boettiger, C, Lloyd, H, Chadès, I. A Shiny r app to solve the problem of when to stop managing or surveying species under imperfect detection. Methods Ecol Evol. 2020; 11: 1707– 1715. https://doi.org/10.1111/2041-210X.13501
Adaptive management or learning by doing, is praised as the best practice method to manage natural systems under uncertainty (see ESA’14 talk). Limited for a long time by our ability to solve adaptive management problems, our research now allows us to find the best adaptive management strategies when networks change over time. This was made possible thanks to our research in Artificial Intelligence (AI) and Conservation science.
What have we discovered in 2 steps?
1) Unlocking the beast. Being strategic about adaptive management means finding the best management strategy when we don’t know exactly what will happen in the future (structural uncertainty). Until very recently, finding the optimal strategy to such decision problems was possible for very small size problems, limiting the application of adaptive management principles. In 2012, we published a fundamental paper that demonstrates that adaptive management problems can be solved using a simplified POMDP (Partially Observable Markov Decision Process, see tiger paper). This is an important finding because modelling an adaptive management problem as a POMDP means we can use very fast algorithms from AI and solve very large adaptive management problems. On a side note, this paper was published at the top AI Conference (AAAI) and received “best paper award” (Computational Sustainability track, thanks for the support!).
2) Boldly go where no one has gone before. Our second step was to demonstrate the power of our findings on the most complex problem we could imagine. Thinking about it, the most difficult problems to solve in ecology are spatial problems (migratory networks) with changing dynamics over time (non stationarity, climate change) for which the consequences on species management are unknown (structural uncertainty, population dynamics). Well, we did it! Check our splendid paper in Proceedings B led by Sam Nicol that brings it all together. This work is amazing for so many good reasons: the shorebird application, the fundamental AI research, the writing, the figures, the authors, the journal and the 20-page supplementary information!
That is now official, I am the new team leader of the conservation decisions team and we will be hiring this year! So stay tuned if you are looking for a wicked postdoc position in adaptive management/computational sustainability!